Recently had quite a few conversations with brands about measurement which eventually leads into attribution. There is a common trend that I have seen with all brands which I share below.
An overarching issue is how channel performance is evaluated which is not directly an attribution issue, but it has a lagging impact when it comes to looking at attribution. For example, Display campaigns are compared to Search campaigns which are not comparable as the objective of both are very different. Which leads into how both Display and Search campaigns are measured, using the same metrics i.e., conversion rate for both is not a fair comparison.
The three common mistakes I have seen brands make when it comes to attribution:
1 – Believing everything your Web (Google) Analytics tells you
Web Analytics is there to understand the traffic that comes to your website and the user journey to taking a certain action i.e., buying a product. The web analytics tool is not there to measure media performance.
Attribution within web analytics is broken as it focuses on anything post click not post impression / view. Channels such as Paid Search and Organic Search will always look good within attribution as both those channels will drive a high volume of traffic.
Channels that are highly targeted with engaging content may not drive an immediate click, but that user will likely visit the site coming from Branded Search or Direct. Using web analytics attribution, the credit will not be given to the channel driving the engaging content, it will be given to Branded Search or Direct.
Believing everything your web analytics attribution tells you is going force you down a rabbit hole. You won’t get the full picture of how your audience are behaving and most likely reward the channels that are stealing the demand created by other channels. Which is why Google love brands using Google Analytics for attribution as it ensures that Paid Search always looks good.
2 – Remove brand search from attribution
Branded paid search provides huge value in the wider marketing mix, but it should not be accounted for within attribution. Branded search provides value in controlling the user journey from the ad-copy to landing page and protect against competitors bidding on the terms.
Any user who types in a brand name into Google has the intention to visit the website.
What this means any channels that helped drive the demand their performance will be weakened, using last click in most cases it will benefit branded search.
To truly understand the value of your channels and to drive better performance, branded paid search needs to be stripped out of attribution.
3 – Re-targeting provides bad data for attribution
The general rule is that most sales would have occurred without the re-targeting campaign. When looking at performance on a last click or post view performance it will always over-credit the re-targeting campaigns.
Google and Facebook are the big winners, with their massive pool sets of data encouraging brands to invest more into re-targeting. Which in turn drives poor management of the campaigns with high frequency and low viewability which annoy the shit out of users.
There is value in re-targeting campaigns but not as much you believe. Run a test and reduce your re-targeting campaign budget say by 50% and monitor the impact it has on sales.
The world of measurement and attribution is complicated, making these changes are steps in the right direction to achieving better outcomes. Points 2 and 3 can be implemented much quicker, point 1 is going to require internal stakeholder backing as its entails additional investment.
